Archive. This is an archive of the Apple Support Community. Discussions are being restored and soon you will be able to read and post here again.

Is Golden Gate Siri an AI Agent?

I'm using Golden Gate 27.2


In the news Canadian banks are starting to ban the use of AI Agents and will not honour losses due to fraud. Meta's Muse was mentioned as an example however we use Apple in these forums. The article said 3rd party agents but is Muse 3rd party?


Is Golden Gate now which updated Apple Intelligence and Siri part of it? I have been using the Apple Keychain in the cloud for as long as I can remember. I have complex passwords and no two are alike.


Does anyone know these days of the Keychain in the cloud is now an Apple Al Agent? If so I doubt it's 3rd party. I will call my bank tomorrow as well.

10 replies

Yes Meta Muse is a 3rd party Agenic AI agent. These are models that will automatically order goods and services for you and possibly when you don't even want them to. I would certainly understand a bank refusing to honor losses due to fraud when using any such model. Meta platforms, such as Facebook, Instagram, etc, are already full of scams, now they are going to add an AI tool that is capable of making its own determination for automatic purchases?


  • No Siri AI is not an agenic agent that can act on its own to make any purchases.
  • No Golden Gate does not include an agenic AI agent that is going to be making purchases.
  • No the Keychain is not an Agenic AI model.
zenon258 wrote:
Thanks so much for this very informative reply. I didn't think the Keychain was an AI model but I was not sure about Siri.

To be clear Siri AI does use an AI model, it is a generative AI model, not an agenic model. You would have to contact your bank to determine what they are referring to because I cannot speak for them. There are 2 types of AI being used, Generative AI and the new form referred to as Agenic AI.

  • Generative AI used by Apple on your device is for creating content from a prompt, such as answering a question or creating images, text, or code from an input.
  • Agenic AI is an autonomous model for creating a task, workflows, and achieving a goal by taking action on its own. This includes the agents such as Muse you had referred to that are being marketed as a shopping tool that can complete transactions using your credentials stored in 1Password on its own. Of course, Meta says they include safeguards, but Meta cannot even prevent scams on their Marketplace platform now. They have a poor track record on protecting users Personal Information and were fined $25 billion this month for intentionally misleading consumers about personal data privacy and allowed third-party access to user information.


From a banks perspective, I could certainly see where these Agenic AI models would be a concern for them. A customer saying they did not make a charge, it was the Agenic AI model they used that did it, is not going to be a valid reason for disputing a charge.

Re: “… and will not honour [replace a customer’s] losses due to fraud …”


Come again ??? 😳


This might seem to be counter to at least some consumer protection laws.


Have you an authoritative source citing any details?

Thanks. Today is the first time I have heard the term agenic. Agentic? Either way it has been a very good eye opener for me. Something you task to do in the future and the AI agent figures out how to handle it. It's pretty clear to me now and I'm glad I asked. Learned something new today.


I have noticed with GG 27.2 and my online bank account. At one time I could click on both the user and password fields and it would enter both. Now I can still click on the password but I have to go into my Keychain for the password. Same on both my Mac mini M4 and MB Air M3.

zenon258 wrote:
https://globalnews.ca/news/12124433/ai-agents-canadian-banks/

To me that makes sense. I think banks are well within their right to block transactions from these AI agents. The customer should not be able to simply claim that they did not make the transaction, it was the AI agent they used that did it. The responsibility should be on the developer of that AI agent that made the transaction. It is up to them to provide the safeguards/guardrails for their technology and in my opinion they are failing miserably.


The question should be, who is responsible for refunding these transactions, when a dispute arises. The banks?

Thanks so much for this very informative reply. I didn't think the Keychain was an AI model but I was not sure about Siri.

Thank you for the reply.

NP.